Does Global Cluster works with only one interface for Application AND Cluster?

[Rudasa]

Hello gurus,

I have setup two dual-node cluster in the following way:

Site 1 with Primary Server (10.31.150.111) and Secondary Server (10.31.150.112)

Site 2 with Primary Server (10.31.150.101) and Secondary Server (10.31.150.102)

I have an application running as Global Group with Virtual IP: 10.31.150.110 (Site 1) and 10.31.150.100 (Site 2)

The application uses Oracle and RVG is implemented for VM Replication (and configured in VCS)

Global Cluster is configured (apparently ok) with Icmp Heartbeat. However, when I check the

status in each cluster, both are telling me that Local Cluster is RUNNING and Remote Cluster is INIT (?)

More than that, the GUI (hagui) tells me "Site x is not running" as remote status.

Is it possible to implement all this having ONLY ONE IP address in each Site for Application and Global Cluster?

Thanks

[Rudasa]

Hello gurus,

Forgot to mention: in this setup I'm explaining, the problem is: I CAN'T MANUALLY SWITCH Between Clusters (GlobalFailOverPolicy is Manual).

I have setup two dual-node cluster in the following way:

Site 1 with Primary Server (10.31.150.111) and Secondary Server (10.31.150.112)

Site 2 with Primary Server (10.31.150.101) and Secondary Server (10.31.150.102)

I have an application running as Global Group with Virtual IP: 10.31.150.110 (Site 1) and 10.31.150.100 (Site 2)

The application uses Oracle and RVG is implemented for VM Replication (and configured in VCS)

Global Cluster is configured (apparently ok) with Icmp Heartbeat. However, when I check the

status in each cluster, both are telling me that Local Cluster is RUNNING and Remote Cluster is INIT (?)

More than that, the GUI (hagui) tells me "Site x is not running" as remote status.

Is it possible to implement all this having ONLY ONE IP address in each Site for Application and Global Cluster?

[TomerG]

The ClusterService group in each cluster should have a WAC resource. Is that started successfully?

BTW, correct, 1 interface is all that's required, since you don't have more than one node in each local cluster. For redundancy, however, it'd be better to use some kind of IP failover mechanism, like IPMP, or IP bonding/trunking, if you have that option, to prevent Network Card single point of failure.